O-015 A Statistical Analysis of Safety Test Results and Implications for Insensitive Munitions
In the last 12 months, two countries, France and the United States, have issued or updated safety policies regarding their munitions or insensitive munitions. At NATO, a Standardization Agreement (STANAG) regarding IM is in its final draft and incorporates standardized testing.
Responses to these tests will help to decide if a munition meets safety and/or IM requirements. Also cost considerations affect why a limited number of munitions may be allocated for a given test.
In a presentation at the DDESB seminar in 1992, NIMIC focused on the poor reproducibility of some of these standardized tests (qualitative aspects), and hence the necessity to couple experimental testing with modeling.
A further study by NIMIC is presented which deals with the probability of information that can be misleading as a result of interpretation of the test responses of a few items, selected from a large production lot.
Through its extended database and the commitment of some of its European points of contact, NIMIC has conducted a statistical study of data involving repetitive bullet impact test reports series on various munitions, e.g. the 155 mm M 107 artillery shell and the General Purpose MK 82 bomb. It has focused on parameters such as:
- error of first kind
- error of second kind
- the operating characteristic curve of the test
The study enabled NIMIC to propose an assessment of the degree of confidence of a test series versus the number of tests conducted. The particular case of the standardized NATO bullet impact test procedure, requiring 2 items to be tested, has been addressed and its poor level of confidence highlighted.
In this paper, some conclusions have been drawn to improve the bullet impact test procedure and its reliability. However, exchanging data on, and applying models to other subscale or similar configurations within the NIMIC countries are advocated in order to replace all-up round testing.